Posts by Tuna Ertemalp

1) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1774212)
Posted 26 Mar 2016 by Profile Tuna Ertemalp
Post:
Bump. Now that virtually all tasks pushed out are v8, the current XML file communicating the server status to external sites is essentially useless without this getting fixed.

Thanks
Tuna
2) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1770283)
Posted 7 Mar 2016 by Profile Tuna Ertemalp
Post:
Bump. ClimatePrediction had a similar issue, which I had reported and they just fixed. I am hoping that SETI will be fixed, too, sooner than later... :)
3) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1767626)
Posted 26 Feb 2016 by Profile Tuna Ertemalp
Post:
Thanks!
4) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1767538)
Posted 25 Feb 2016 by Profile Tuna Ertemalp
Post:
So,

Human readable forms are on http://setiathome.berkeley.edu/sah_status.html for main and on http://setiweb.ssl.berkeley.edu/beta/server_status.php for beta, by just following the links on the project sites.

For beta, your pointer to http://setiweb.ssl.berkeley.edu/beta/server_status.php?xml=1 seems to reflect the numbers on the human readable page. Thanks!

For main, however, BOINCstats is looking at http://setiathome.berkeley.edu/sah_status.xml and you say http://setiweb.ssl.berkeley.edu/sah_status.xml is that of main, and they both seem to be the same file anyways, and they are both NOT reflecting the human readable data. Doing something like http://setiweb.ssl.berkeley.edu/server_status.php?xml=1 returns 'Server Error 500'.
5) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1767521)
Posted 25 Feb 2016 by Profile Tuna Ertemalp
Post:
And, for SETI@Home Beta, assuming http://setiweb.ssl.berkeley.edu/sah_status.xml is supposed to match http://setiweb.ssl.berkeley.edu/beta/server_status.php, well, they don't, leading to a similar problem on BOINCstats for that project, also. Even if that is not the right XML I am looking at, BOINCstats has the same issue with SETI@Home Beta, so there must be an XML not matching the Server Status page.

Tuna
6) Message boards : Number crunching : Server Status page & XML don't agree; v8 tasks are missing from the XML (Message 1767518)
Posted 25 Feb 2016 by Profile Tuna Ertemalp
Post:
Summary: It seems the XML server status report is missing the V8 tasks. Therefore, BOINCstats shows wrong data for project status for SETI@home.

Details: Right now, http://boincstats.com/en/page/projectStatusInfo thinks that SETI@home has only less than a handful tasks ready to be sent out and 20x less tasks in progress, as opposed to what http://setiathome.berkeley.edu/sah_status.html shows. Turns out, BOINCstats reads http://setiathome.berkeley.edu/sah_status.xml and that XML doesn't seem to have the right numbers:

<sah_results_ready_to_send>2</sah_results_ready_to_send>
<ap_results_ready_to_send>0</ap_results_ready_to_send>
<sah_v8_results_ready_to_send>0</sah_v8_results_ready_to_send>
. . .
<sah_results_in_progress>38446</sah_results_in_progress>
<ap_results_in_progress>150109</ap_results_in_progress>
<sah_v8_results_in_progress>150109</sah_v8_results_in_progress>


compared to what the human readable page shows:

Data Distribution State     SETI@home v7 #     Astropulse #     SETI@home v8 #     As of*
Results ready to send               2                   0            540,576          0m
Results out in the field       38,446             150,109          4,089,293          0m


Notice how the two V8 lines simply repeat the data in the preceding AP lines. If you inspect the rest of the XML from the link above, you'll notice that is the pattern for EVERY V8 entry, so there is a much greater issue here than just those two numbers.

The original thread on BOINCstats: https://boincstats.com/en/forum/5/10756,1#140622

Thanks
Tuna
7) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1761400)
Posted 1 Feb 2016 by Profile Tuna Ertemalp
Post:
May I hope all the CPU and GPU tasks that come back with non-abort errors or as invalid, along with stack traces I see when I drill into them, are aggregated into a top-level developer's view and analyzed for problem trends, resulting in fixes incorporated into future releases, instead of depending on individuals reporting them?

Tuna
8) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1761070)
Posted 31 Jan 2016 by Profile Tuna Ertemalp
Post:
Specifically the thread: Optimised Applications and Other Binaries if any has trouble finding it.

Yesssss.... Thanks! :-)
9) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1761054)
Posted 31 Jan 2016 by Profile Tuna Ertemalp
Post:
Is there a main Lunatics announcement thread (as opposed to per version threads) that I can subscribe to so that I can hear about future releases, instead of checking the whole Number Crunching forum continuously?

?
10) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1760671)
Posted 30 Jan 2016 by Profile Tuna Ertemalp
Post:
Keep in mind that GBT tasks are different.

GBT?

Green Bank Telescope

News


Nice. Can/will one be able to tell if a task is from GBT, like one can tell VLARs since task name has .vlar_n at the end?
11) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1760652)
Posted 30 Jan 2016 by Profile Tuna Ertemalp
Post:
Keep in mind that GBT tasks are different.


GBT?
12) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1760651)
Posted 30 Jan 2016 by Profile Tuna Ertemalp
Post:
My results are coming in (http://setiathome.berkeley.edu/results.php?hostid=5551867). Look at the validation pending or valid ones. In the past, with stock apps, it was about 600-750secs runtime per task, i.e within your expectation (that is about 5-6 WUs per hour). Now, with me pushing 9 tasks at a time with Lunatics app, it went up to 2100-2600 secs per task (that is about 12-15 WUs per hour). So, overall a gain, but I wonder what would have been if I were to push multiples with the stock app or no multiples with the Lunatics app. I guess I'll have to do some tests in the morning after seeing what this does over night.

As comparison:
- This is a non-Lunatics machine with Dual Titan X: http://setiathome.berkeley.edu/results.php?hostid=7916726
- This is a non-Lunatics machine with Quad Titan X: http://setiathome.berkeley.edu/results.php?hostid=7856578

Tuna

EDIT: Actually, the times jumped up to 4000+ per WU, reducing the per hour to 7ish. Not good. 2000+ ones must have started when I had less parallelism, thus were faster at their start. I'll reduce to 0.5 or 0.3 until morning.
13) Message boards : Number crunching : Lunatics Windows Installer v0.44 - new release for v8 (required upgrade) (Message 1760632)
Posted 30 Jan 2016 by Profile Tuna Ertemalp
Post:
Maybe a dumb question, but first time I installed Lunatics on any of my machines. So... Bare with me.

Machine has an EVGA Titan X Hybrid (and an old i7-970 CPU). I see the GPU task running: 0.04 CPU + 1 GPU. Using Lunatics_x41zi_win32_cuda50.exe.

Then I look at MSI Afterburner, and I see the GPU use to be very low: Usage% is usually <10% with a few peaks into 30-60% range, but just momentary peaks. Power% is at about 50%. Memory use is at just 700MB out of available 12G. This is at 41mins in with 17mins estimated left. Is this normal? Feels like GPU is extremely underused.

Also, during Setup, for the CUDA, the installer defaulted to the third option for the 2xx cards, not the first CUDA5 option for the Kepler/Maxwell cards. Doesn't the Setup detect the card(s) installed and suggest the right thing, which it seems to be doing for the CPUs?

And, reading stuff below: What is <count> for GPU and CPU? Am I supposed to set <count> to 0.3 to get three tasks run on one GPU? If so, in app_info or app_config? In general, is there a document somewhere explaining the format for the app_info and app_config?

And, finally: Is there a main Lunatics announcement thread (as opposed to per version threads) that I can subscribe to so that I can hear about future releases, instead of checking the whole Number Crunching forum continuously?

Tuna

EDIT: Per Grant's example, I tried throwing 3, 5, 10, 8, 9 tasks to the GPU, and at 9 tasks, I get 75-90% usage (10 tasks made it stuck at 100%), 65% power usage, 3230MB memory use. But this is just me blindly following Grant';s template & Afterburner graphs. Is there a downside to this? The GPU is at a nice 39C. I used this:
<app_config>
  <app>
    <name>setiathome_v8</name>
    <gpu_versions>
      <gpu_usage>0.11</gpu_usage>
      <cpu_usage>0.04</cpu_usage>
    </gpu_versions>
  </app>
</app_config>
14) Message boards : Number crunching : New rig project (Message 1759601)
Posted 27 Jan 2016 by Profile Tuna Ertemalp
Post:
I have recently built two machines, one with Quad Titan X, one with Dual Titan Z (specifically for PrimeGrid given the Double Precision prowess of the Z). Both 24/7 BOINC machines. So, I'll tell you what I used in the Quad X:

ASUS X99-E WS/USB 3.1 Motherboard. 4-way, all Gen3 x16, M.2 doesn't share anything with PCI, nor is limited by anything.

SSD 950 PRO M.2 512G boot drive. I am extremely happy. For a BOINC-only machine, that should be enough storage for OS+BOINC Data and way more.

SSD 850 PRO 1T SSDs, if you need more storage, as many as you need (board allows up to 12), and then use Windows 10 Storage Spaces to set them up as parity or striping. Don't use the onboard RAID since you'll lose the TRIM functionality on the SSD. These guys are beasts, never had one fail, they are in all of my machines.

Fo optical, I used LG Internal SATA 16x Super Multi Blue WH16NS40 and put it into an external enclosure (VANTEC ST-530S3-BK Aluminum 5.25" SATA USB 3.0 External 5.25" Optical Drive Enclosure USB 3.0 External Interface). Why external? Read on to see my comments about airflow.

For GPUs, I prefer EVGA. I get them individually sealed water cooled, resulting in as many radiators as GPUs. For EVGA, that is either 'EVGA GeForce GTX TITAN X HYBRID GAMING' or 'EVGA GeForce GTX TITAN X SC GAMING' combined with 'EVGA All-in-One Hybrid Water Cooler for Select EVGA GTX TITAN X Graphics Cards' plus some elbow grease.

For CPU, I love Intel® Core™ i7-5960X Processor Extreme Edition (20M Cache, up to 3.50 GHz). 8 cores, 16 threads, and I was able to push it to 3.9GHz.

To cool that, I use CORSAIR Hydro Series H110i GT Extreme Performance Water/Liquid CPU Cooler.

To accommodate all these 4+2=6 fans to exhaust, plus accommodate further intake and exhaust fans, I use Thermaltake Core x9 case. It is big, but has lots of space for anything you might want to do in there. I get two 200 fans as intake for the front, a 140 in the back for exhaust, plus the 4+2=6 fans exhausting at the top. Also the GPUs end up exhausting in the back with their on-card fans in addition to the liquid loop exhausting at the top of the case. And, the PSU also sucks in internal air and exhausts in the back. I removed as much of the stuff as possible inside, like many of the harddisk cages, put the SSDs at the bottom of the case, all in the name of unblocking the airflow in the case, especially that from the front intakes.

I maxed the RAM at 64G (max for that CPU) to be able to use "Keep projects in memory while switching" since I churn for 50 projects. What works with that MoBo and is really good and can be overclocked by the ASUS board automatically: Corsair Dominator® Platinum Series 64GB (8 x 8GB) DDR4 DRAM 2400MHz C14 Memory Kit (CMD64GX4M8A2400C14)

Here is the twist: When I build such an expensive top gear, I want to put it on a UPS, something from APC, to avoid hurting it through flickers and brownouts. Unfortunately, at max CPU+GPU use this beast draws about 1.2KwH. The 1500VA UPS units max out at about 860w. So, I would have needed to get a 2200VA unit which are really expensive and incredibly heavy (i know, because I got one for the dual Titan Z rig, and my jaw fell to the floor when I saw the huge wooden crate FedEx delivered). So, instead, given the case has spots for two PSUs, I got one Antec HCP-1300 Platinum 1300W and one Antec HCP-1000 Platinum 1000W. I put GPUs on 1300w, everything else (MoBo, fans, SSDs, etc.) on 1000. Then I was able to get two 1500VA UPS boxes, put one between the wall and each PSU; cheaper, lighter, plenty room. Both UPS show about 600-650w draw, well below their peak capacity, and well balanced. And, I used Antec since they come with an OC-link cable to link both PSUs so that your On/Off button on the case will turn on/off both PSUs at the same time, and do other things that two PSUs need to agree on before things on your motherboard start acting up due to getting powered by different PSUs. In other words, you connect them, and you have 2300w; the rest just works. When I first had them on one EVGA 1600 (which I later moved to the Titan Z machine), even though the math works out, with all the fans & overclocking, GPUs were getting limited by power/voltage etc. Now, they have all the juice they need even when they are pushing it.

I am also using MSI Afterburner 4.2 software on the machines to watch the GPU and CPU temps, as well as create custom on-GPU fan profiles. I don't actively push for overclocking. At stock, they are plenty fast. But, this whole hardware sort of auto-overclocks until it reaches some thermal/power limit. When that happens, MoBo and GPU fans become important. The factory settings are such that GPU fans don't really exceed 40%. That is nothing when the heat reaches 80C and your cards are thermal limited. In another of my Dual Titan X machines that is not liquid cooled, just by applying the fan profile to kick in high gear at 60-70C, the machine is now running at 50-60C instead of 80-85C, getting more calculation cycles done. Of course, it is louder, and the room is much warmer. :)

Why did I use Titan X instead of 980 Ti? First, just wanted to. Second, and more importantly, wanted this machine to be future proof in case some project started really using the GPU memory, given that I churn for all BOINC projects, including ones that'll come along in the future. So far, that was a waste of money. Max I saw in MSI Afterburner was about 2-3GB use of the GPU memory by some project; probably PrimeGrid. So, 980 Ti with 6G memory should be more than enough, and save you bunch of money, taking the sting away.

Side note: Titan X is able to do 4K with 4:4:4. A good "monitor" to take full advantage of that is actually a TV! I got a Samsung UN40JU6700 Curved 40-Inch 4K Ultra HD Smart LED TV (2015 Model) from Amazon which was an amazing deal in December at $650, including a BluRay player to boot! I am very very happy with it.

Good luck!
Tuna

Disclaimer: No, I am not a hardware geek. If you start drilling into me with detailed questions, I might not be able to say more than above. But, I have a friend who really knows this stuff and enjoys building things for his friends, so I learned quite a bit from him over the last few months. Still, my knowledge is very superficial compared to what he knows. A lot of the above is him holding my hand through the hardware jungle...
15) Message boards : Number crunching : Reinstall/rename of a host caused "Unable to handle request: No such computer", inability to merge, possibly lost credit (Message 1759584)
Posted 27 Jan 2016 by Profile Tuna Ertemalp
Post:
Seems all fine, now. I can see the new identity of the host. Hopefully the credits are correct. Without having recorded it before, I have no way of verifying...
16) Message boards : Number crunching : Reinstall/rename of a host caused "Unable to handle request: No such computer", inability to merge, possibly lost credit (Message 1759318)
Posted 26 Jan 2016 by Profile Tuna Ertemalp
Post:
It might be running, but if you look to the right you'll see that it's still offline. ;-)


Ah! Live'n'learn... :)

Thanks a bundle!
17) Message boards : Number crunching : Reinstall/rename of a host caused "Unable to handle request: No such computer", inability to merge, possibly lost credit (Message 1759314)
Posted 26 Jan 2016 by Profile Tuna Ertemalp
Post:
"BOINC replica database carolyn Running"

And, still the same problem. Next step?

Tuna
18) Message boards : Number crunching : Reinstall/rename of a host caused "Unable to handle request: No such computer", inability to merge, possibly lost credit (Message 1759210)
Posted 25 Jan 2016 by Profile Tuna Ertemalp
Post:
(Interestingly enough, just when I clicked POST for this message last night, around Sun Jan 24 9:15pm, I found the whole SETI website not accessible even though it was up just a few minutes earlier. Seems there was something bad crawling around at the time I was trying to do the below stuff, and might be the reason why a productive machine of mine is now in limbo.)

I had this host EXTBEDROOM2 with lots of credits on SETI@Home and SETI@Home Beta along with ~50 other projects. I had to reinstall a different build of the OS on it from scratch, and while at it, changed its name to EXTBEDROOM. Once this was done, in all 50 projects EXTBEDROOM and EXTBEDROOM2 looked like two separate hosts, so I merged the "old" EXTBEDROOM2 into the "new" EXTBEDROOM in all of them by clicking on EXTBEDROOM's Details link, choosing Merge, picking EXTBEDROOM2 from the suggested compatible host list (*NOT* SelectAll since there are other hosts I have with the same specs) and completing the merge...

...except in SETI@Home and SETI@Home Beta. In both projects, clicking on EXTBEDROOM's Details link gives "Unable to handle request: No such computer". Therefore, I am posting this same thing on both forums, not knowing if it is the same set of people in charge of the database.

Stupidly, in SETI@Home, since I was not able to go into the Details of EXTBEDROOM, I went into the Details of EXTBEDROOM2 (the "old" identity of the host), told it to Merge, which suggested EXTBEDROOM, I accepted it, and now I am seeing only EXTBEDROOM, with very small amount of credits, presumably earned since I reinstalled it, and I still cannot get into its Details page. And, I did not make a note of the HostID of EXTBEDROOM2 before it disappeared not knowing that this would happen.

In short,

* In SETI@Home, EXTBEDROOM2 (hostID=?) should have been merged into EXTBEDROOM (hostID=7916726), but now only EXTBEDROOM exists with small credit, and is inaccessible
* In SETI@Home Beta, EXTBEDROOM2 (hostID=75957) should be merged into EXTBEDROOM (hostID=77934), but EXTBEDROOM is inaccessible.
* Note: Earlier in the day, another smaller/older/weaker machine named EXTBEDROOM was reinstalled, physically moved and renamed successfully across all 50 projects to GARAGE (hostID=7915171 for SETI, 77898 for SETI Beta), and it shouldn't be confused with any of the above.

Is there anything that can be done? My attempts to play with this stuff happened roughly between Sun Jan 24 4-9pm PST, in case logs need to be looked at.

I certainly hope I didn't lose the EXTBEDROOM2's SETI@Home and SETI@Home Beta credits... That'd be a huge bummer.

Thanks
Tuna





 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.